Repairing confusion and bias errors for DNN-based image classifiers.
Yuchi TianPublished in: ESEC/SIGSOFT FSE (2020)
Keyphrases
- single image
- input image
- image representation
- image analysis
- image features
- image data
- image content
- multiscale
- image pixels
- training data
- image distortion
- image collections
- template matching
- region of interest
- image matching
- image segmentation
- image classification
- image retrieval
- naive bayes
- image regions
- spatial information
- feature values
- support vector
- segmentation method
- improve recognition accuracy
- segmentation algorithm
- camera movement
- classifier training
- object detectors
- pixel values
- high resolution
- low level
- edge detection
- hough transform
- keypoints
- image quality
- gray level